Osteogenesis of Heterotopically Transplanted Mesenchymal Stromal Cells in Rat Models of Chronic Kidney Disease


  • Rafael Kramann,

    Corresponding author
    1. Division of Nephrology and Clinical Immunology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    2. Institute of Pathology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    • Address correspondence to: Rafael Kramann, MD, Department of Nephrology and Clinical Immunology, Medical Faculty RWTH Aachen University, Pauwelsstraße 30, 52074 Aachen, Germany. E-mail: rkramann@gmx.net

    Search for more papers by this author
  • Uta Kunter,

    1. Division of Nephrology and Clinical Immunology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    Search for more papers by this author
  • Vincent M Brandenburg,

    1. Division of Cardiology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    Search for more papers by this author
  • Isabelle Leisten,

    1. Institute of Pathology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    Search for more papers by this author
  • Josef Ehling,

    1. Institute of Pathology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    2. Department of Experimental Molecular Imaging (ExMI), Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    Search for more papers by this author
  • Barbara M Klinkhammer,

    1. Division of Nephrology and Clinical Immunology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    Search for more papers by this author
  • Ruth Knüchel,

    1. Institute of Pathology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    Search for more papers by this author
  • Jürgen Floege,

    1. Division of Nephrology and Clinical Immunology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    Search for more papers by this author
  • Rebekka K Schneider

    1. Institute of Pathology, Medical Faculty Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, Aachen, Germany
    Search for more papers by this author


The current study is based on the hypothesis of mesenchymal stromal cells (MSCs) contributing to soft-tissue calcification and ectopic osteogenesis in chronic kidney disease (CKD). Rat MSCs were transplanted intraperitoneally in an established three-dimensional collagen-based model in healthy control animals and two rat models of CKD and vascular calcification: (1) 5/6 nephrectomy + high phosphorus diet; and (2) adenine nephropathy. As internal controls, collagen gels without MSCs were transplanted in the same animals. After 4 and 8 weeks, MSCs were still detectable and proliferating in the collagen gels (fluorescence-activated cell sorting [FACS] analysis and confocal microscopy after fluorescence labeling of the cells). Aortas and MSC-containing collagen gels in CKD animals showed distinct similarities in calcification (micro–computed tomography [µCT], energy-dispersive X-ray [EDX] analysis, calcium content), induction of osteogenic markers, (ie, bone morphogenic protein 2 [BMP-2], Runt related transcription factor 2 [Runx2], alkaline phosphatase [ALP]), upregulation of the osteocytic marker sclerostin and extracellular matrix remodeling with increased expression of osteopontin, collagen I/III/IV, fibronectin, and laminin. Calcification, osteogenesis, and matrix remodeling were never observed in healthy control animals and non-MSC–containing collagen gels in all groups. Paul Karl Horan 26 (PKH-26)-labeled, 3G5-positive MSCs expressed Runx2 and sclerostin in CKD animals whereas PKH-26-negative migrated cells did not express osteogenic markers. In conclusion, heterotopically implanted MSCs undergo osteogenic differentiation in rat models of CKD-induced vascular calcification, supporting our hypothesis of MSCs as possible players in heterotopic calcification processes of CKD patients. © 2013 American Society for Bone and Mineral Research.


Cardiovascular disease accounts for more than one-half of all deaths in patients with end-stage renal disease (ESRD).[1] Vascular calcification is an important contributor to this cardiovascular mortality.[2] Once considered a passive process, it has now emerged that vascular calcification is a tightly regulated, coordinated process resembling osteogenesis.[3-5] Indeed, about 15% of atherosclerotic human vessels contain completely trabecular bone with all stages of enchondral ossification and even marrow cavities with hematopoietic cells.[6-9] The appearance of new bone formation inside the wall of arteries suggests that osteoblast progenitors must be recruited during the pathogenesis of vascular calcification.

Nevertheless, it remains highly controversial whether osteoblasts are derived from resident cells such as vascular smooth muscle cells (VSMCs) or pericytes, or whether they derive from circulating progenitor cells.[10] It has recently been appreciated that mesenchymal stromal cells (MSCs) are pericytes.[11] Because all cells involved in enchondral ossification share a common mesenchymal progenitor cell, we hypothesize that MSCs, in their role as pericytes, contribute to ectopic ossification by differentiating into osteogenic effector cells during vascular and soft-tissue calcification. In a recent study we established a three-dimensional cultivation system containing the major fibrillar collagens of the vessel wall—collagen I and III—in combination with MSCs, mimicking the vascular MSC niche.[12] Furthermore, we demonstrated that uremic serum–exposed human bone marrow (BM)-MSCs differentiate into functional osteoblasts and were able to calcify this three-dimensional vascular collagen matrix.[12]

However, without a distinct marker it is difficult to analyze the role of MSCs during pathophysiologic processes in their in vivo niche.[13, 14] Thus, we now established a heterotopic vascular niche to study the calcification potential of MSCs in a chronic kidney disease (CKD) setting in vivo. We implanted these three-dimensional collagen gels containing rat BM-MSC intraperitoneally in two rat models of CKD with vascular calcification: (1) 5/6 nephrectomy plus high phosphorus diet (5/6 Nx), and (2) adenine nephropathy (AdN) in comparison to healthy control animals. The experiment is based on the hypothesis that MSCs only calcify their surrounding collagen I/III matrix in the environment of CKD. Our implantation model allows a systematic comparison of MSC-driven extracellular matrix (ECM) remodeling in the implanted heterotopic collagenous vessel-like MSC niche to the ECM remodeling that occurs during vascular calcification in the vessel wall in CKD.

Materials and Methods


Local government authorities (the Landesamt für Natur Umwelt und Verbraucherschutz [LANUV], Nordrheinwestfalen [NRW]) approved all animal experiments reported in this article. Animals were housed under standard conditions with constant temperature and humidity, 12-hour light/dark cycles, and with free access to drinking water. We used inbred male Wistar rats weighing 290 to 320 g (Charles River, Erkrath, Germany).

Isolation and cultivation of MSCs

MSCs were isolated from the bone marrow of male Wistar rats weighing 180 to 200 g (Charles River, Erkrath, Germany) and cultivated and differentiated according to previously described standard protocols.[15] All MSCs fulfilled the minimal criteria for MSCs; ie, they were plastic adherent when maintained in standard culture conditions, differentiable in adipocytes and osteoblasts, and had a typical marker pattern of CD44 + , CD73 + , CD90 + , CD45–, CD34–, and CD31–(Supplementary Fig. 1). Cells were expanded and used for experiments in passage 3. Three-dimensional collagen gels containing MSCs were generated as described, with pooled MSCs from 9 donors (1 gel volume = 500 µL, containing 5 × 106 MSCs).[12, 16, 17] After 7 days of cultivation the collagen gels were implanted.

Experimental models and experimental design

The experimental schedule is shown in Fig. 1A. Two different models for the induction of CKD with vascular calcification in rats were used:

  • Model 1. 5/6 Nx was induced in a one-step surgery: rats were anesthetized, a 6-cm incision was made in the linea alba, and after nephrectomy of the right kidney, the left kidney artery was exposed. Two of the normally three-segment arteries were ligated (7-0 suture, monofil; Ethicon, Livingston, UK). To aggravate the vascular calcification in this model the 5/6 Nx rats were fed a high phosphorus (1.2%), normal calcium diet (Altromin, Lage, Germany) as described.[18]
  • Model 2. Experimental AdN was induced by feeding a diet containing 0.75% adenine, 2.5% protein, 1.06% calcium, and 0.92% phosphorus (Altromin, Lage, Germany) over 4 weeks, as described.[19]
Figure 1.

Experimental design and renal function parameters. (A) Experimental time schedule, (B) transplantation areas of collagen gels, (C) macroscopic view of heterotopically transplanted collagen gels after 8 weeks (control group), (D) HE staining of MSC-containing collagen gel, (E) HE staining of non-MSC–containing collagen gel, (FH) clinical renal function parameters over a time period of 8 weeks. Scale bars = 300 µm (D, E), 25 µm (D, E, inserts). *p < 0.05, **p < 0.001 versus control group, b = level of significance valid for both CKD groups. HE = hematoxylin and eosin.

We implanted four MSC-containing collagen gels (one with Paul Karl Horan 26 [PKH-26]-labeled MSCs) around the right kidney bed and four cell-free collagen gels as internal controls around the left kidney bed (Fig. 1BE). The gels were fixed with 9-0 suture (polyamid, monofil; Ethicon, Livingston, UK) in the perirenal adipose tissue (Fig. 1C). We performed a preliminary experiment to ascertain the optimal time point of gel implantation, when the two CKD groups have about the same renal function parameters. Thus, we implanted the collagen gels during the nephrectomy surgery in the 5/6 Nx group and 2 weeks after starting the adenine diet in the AdN group (Fig. 1A). As controls, we implanted the same number of cell-free and MSC-containing collagen gels in weight-and age-matched healthy Wistar rats with standard chow (Ssniff, Soest, Germany; R/M-H ™ 1% calcium, 0.7% phosphorus). Serum samples were collected weekly from the tail vein. Rats were placed in metabolic cages for overnight urine collection 2, 4, 6, and 8 weeks after implantation of collagen gels. Surviving animals were killed by exsanguination (puncture of the vena cava) 4 weeks (control, n = 6; AdN, n = 6; 5/6 Nx, n = 5) and 8 weeks (control, n = 6; AdN, n = 7; 5/6 Nx, n = 7) after the implantation of collagen gels. Note: the 8-week AdN group were fed an additional week of the adenine-diet 6 weeks after implantation of the collagen gels. Anesthesia in all surgical procedures was carried out with xylazine/ketamine intraperitoneally (90/10 mg/kg), animals were treated with 0.05 to 0.1 mg/kg buprenorphine every 12 hours over 3 days after surgery.

Miscellaneous measurements

Serum and urine biochemistry were performed by an autoanalyzer. Rat intact parathyroid hormone (iPTH; Immutopics, San Clemente, CA, USA) and rat bone morphogenic protein 2 (BMP-2) ELISA (Boster Biological Technology, Fremont, CA, USA) kits were used according to the manufacturers' instructions. Calcium was extracted from the collagen gels and aortic sections overnight with HCl (0.6 M) and quantified using cresolphthalein complexone chemistry for a colorimetric assay at 578 nm (Randox Laboratories Ltd., Crumlin, UK). The calcium content was normalized to the tissue dry weight, allowing comparisons.

Fluorescence labeling and fluorescence-activated cell sorting analysis of MSCs

To determine both cell division history and the remaining amount of MSCs in the collagen gels, MSCs of one collagen gel in each animal were labeled with PKH-26 red fluorescence cell linker kit (Sigma-Aldrich, Saint Louis, MO) according to the manufacturer's instructions. The collagen gel containing PKH-26–labeled MSCs was always implanted above the right kidney bed, resulting in a clear detection of this collagen gel after killing. After explantation, this collagen gel was explanted and digested with collagenase I (Biochrom, Berlin, Germany) (after taking a small piece about 20% for histology) at 37°C for 30 minutes. MSCs were stained using primary antibodies against 3G5 (mouse monoclonal, 1:100; Abcam, Cambridge, UK; ab3095), CD31 (clone TLD-3A12; Serotec), rat CD34 (ICO115; Santa Cruz Biotechnology, Santa Cruz, CA, USA), CD44 (MCA643GA; Serotec), rat CD73 (BD Pharmingen, Heidelberg, Germany), rat CD90 (clone OX7; Mediagnost, Reutlingen, Germany), and a goat polyclonal antibody against rat CD45 (M-20; Santa Cruz Biotechnology) followed by staining with according secondary fluorescein isothiocyanate (FITC)-conjugated antibody (Jackson, West Grove, PA, USA). Cells were resuspended in PBS and analyzed by flow cytometry (fluorescence-activated cell sorting [FACS] Canto II; BD Biosciences) within 1 hour; data were analyzed using Flow Jo software (version 7.5; Tree Star Inc, Ashland, OR, USA).

Histomorphological analysis

For histological and immunohistochemical analyses, tissue sections were fixed in 3.7% formaldehyde for 24 hours, paraffin-embedded, cut with a rotating microtome at 3 µm thickness (Leica), and stained according to routine histology protocols. Immunohistochemical analysis was performed using primary antibodies for fibronectin (rabbit polyclonal, 1:50; Millipore, Billerica, MA, USA), collagen I (goat polyclonal, 1:100; Southern Biotec, Birmingham, AL, USA), collagen III (goat polyclonal, 1:100; Southern Biotec), collagen IV (goat polyclonal, 1:200; Southern Biotec), laminin (mouse monoclonal, 1:1000; Sigma), sclerostin (goat polyclonal, 1:200; R&D Systems, Minneapolis, MN, USA), BMP-2 (rabbit polyclonal, 1:200; Abcam, Cambridge, UK), and osteopontin (OPN) (mouse monoclonal, 1:50; Novus Biologicals, Herford, Germany). Control stainings with isotype immunoglobulin G (IgG) or serum (Vector Laboratories, Burlingame, CA, USA) were performed to exclude unspecific absorption of antibodies to the tissue (data not shown).

For detection of PKH-26–labeled MSCs, collagen gels of representative animals were frozen in liquid nitrogen. Cryosections of 4 µm were stained with the following antibodies: sclerostin (rabbit polyclonal, 1:200; Santa Cruz Biotechnology; sc-130258), runx2 (goat polyclonal, 1:50; Santa Cruz Biotechnology; sc-8566), myeloperoxidase (rabbit polyclonal 1:200; Abcam), stromal cell derived factor 1 (SDF-1, rabbit polyclonal, 1:200; Abcam), CD31 (rabbit polyclonal, 1:100; Abcam), Ki67 (rabbit monoclonal; Vector Laboratories), 3G5 (mouse monoclonal, 1:50; Abcam; ab3095), followed after washing by labeling with donkey anti-goat-Cy5, donkey anti-rabbit Cy5, donkey anti-mouse Cy5, cow anti-mouse AF488, respectively (Jackson). After staining with diamidino-2-phenylindole (DAPI; Thermo Scientific, Rockford, IL, USA) sections were mounted in anti-fade Gold (Vector Laboratories). Immunofluorescence staining of the aorta was performed in paraffin-embedded sections after deparaffinization (no antigen demarcation step) using the above mentioned antibodies for 3G5 and sclerostin followed by donkey anti-mouse AF488, donkey anti-rabbit Cy5 (Jackson), respectively. Images were generated using the Nikon C1 D-Eclipse confocal microscope. Scanning electron microscopy and electron dispersive X-ray (EDX) analysis were carried out as described.[12] Protein expression was quantified by two blinded investigators (RK, RKS) using a semiquantitative scoring system (0, no expression; 1, weak expression; 2 moderate expression; 3, strong expression; 4, very strong expression) at a magnification of ×200, as described.[20, 21]

Real-time RT-PCR

Real-time PCR was carried out as described.[12] TaqMan primers and probes were designed from sequences in the GenBank database using the Primer Express software (Applied Biosystems, Foster City, CA, USA). Primers are listed in Supplementary Table 1.

Ex vivo micro–computer tomography

Eight weeks after collagen gel implantation, 2 animals per group were euthanized and ex vivo scanned using the gantry-based dual energy micro–computed tomography (µCT) scanner TomoScope 30s Duo (CT Imaging, Erlangen, Germany). For visualization and quantification of radio-opaque calcifications, a dual-energy scan at 41 kV and 65 kV (at 0.5 mA and 1 mA) was performed for each animal, acquiring 2880 projections over 6 minutes of continuous rotation for each tube. A Feldkamp-type reconstruction (CT Imaging, Erlangen, Germany) was performed with voxel size of 35 × 35 × 35 µm3. After reconstruction, the data generated by µCT were visualized and analyzed with Imalytics research workstation 3.0 beta (Philips Research, Aachen, Germany).

Statistical analysis

Data are presented as means ± SD. Data analyses were performed using Student's t test or Mann-Whitney U test where appropriate. For multiple group comparison, analysis of variance with post hoc Bonferroni correction was applied. Statistical significance was defined as p < 0.05. Analyses were performed using PASW Statistic 18.0 (SPSS Inc., Chicago, IL, USA).


Induction of CKD in rat models of 5/6 Nx and AdN

The experimental schedule is shown in Fig. 1A. Following the establishment of AdN and 5/6 Nx, MSC-containing collagen gels and cell-free collagen gels as internal controls were transplanted around the kidneys as shown in Fig. 1B, C. Collagen gels with MSCs removed after 4 and 8 weeks contained spindle-shaped cells, exhibiting typical MSC morphology (Fig. 1D). Cell-free collagen gels showed migration of cells into the collagen gels, predominantly at the peripheral border, but significantly fewer cells in comparison to MSC-containing gels (Fig. 1E). Both models of CKD led to increased serum creatinine and urea, as well as a significantly reduced creatinine clearance compared to the control animals (Fig. 1FH).

Calcification of MSC-containing collagen gels at 8 weeks in both CKD models

Calcification of transplanted MSC-containing collagen gels (+MSC) was revealed in both CKD models by von Kossa staining (Fig. 2A), EDX analysis (Fig. 2B), and calcium measurement (Fig. 2C), whereas no calcification occurred in the control group or in implanted cell-free collagen gels (–MSC) of all groups.

Figure 2.

Calcification of the aortic wall and the implanted MSC-containing collagen gels in both rat models of chronic kidney disease. (A) Von Kossa staining highlighted calcified areas in the aortic wall and in MSC-containing collagen gels (+MSC) under chronic kidney disease conditions. No calcification was detected in control animals and non-MSC–containing collagen gels (–MSC). (B) EDX analysis identified the nature of the von Kossa–positive deposits in the aortic wall and MSC-containing collagen gels as Ca/P crystals. (C) Calcium content standardized for tissue dry weight shows significantly increased calcium in the aortic wall and the +MSC collagen gels. (D) Weekly serum calcium and phosphate levels measured over the experimental procedure of 8 weeks. (E) Computer tomography images clearly showed the calcification of the vasculature (white arrowheads) accompanied by calcification of +MSC collagen gels (right side, arrows) in chronic kidney disease. (# = calcified remaining kidney part) *p < 0.05, **p < 0.001, b = level of significance valid for both CKD groups. Scale bars = 300 µm (A, upper panel), 50 µm (A, medium and lower panel).

EDX analysis identified the von Kossa–positive deposits in aortas and MSC-containing collagen gels as Ca/P crystals (Fig. 2B; spots of measurement are shown in Supplementary Fig. 2A). Further, the calcium content after 8 weeks was significantly enhanced in MSC-containing collagen gels in comparison to the control animals and cell-free collagen gels (Fig. 2C). At the earlier time point, ie, 4 weeks, we observed a nonsignificant trend toward elevated aortic and MSC-containing gel calcium content in CKD rats (Supplementary Fig. 2). Serum analyses showed significantly decreased calcium levels and increased phosphate levels in both CKD rat models (Fig. 2D).

Finally, dual-energy ex vivo µCT imaging confirmed calcification of the vasculature (Fig. 2E, white arrowheads) and the MSC-containing collagen gels (Fig. 2E, right side, arrows) in CKD rats. In 5/6 Nx the strong vascular calcification was accompanied by enormous kidney calcification (Fig. 2E, rhomb).

Implanted PKH-26–labeled MSCs but no migrated cells expressed markers of osteogenic differentiation

MSCs labeled with the fluorescence dye PKH-26 (red) prior to implantation expressed the osteocyte marker sclerostin (Fig. 3B, arrows) and the major osteogenic transcription factor Runx2 (Fig. 3C, arrows) in both CKD rat models, whereas no expression of Runx2 or sclerostin was observed in the control animals (Fig. 3A) or in migrated PKH-26–negative cells (Fig. 3B, C, arrowheads). These findings indicate that the implanted MSCs are driving cells in the calcification of implanted collagen gels and undergo osteogenic differentiation. After 8 weeks the majority of cells in collagen gels were still positive for PKH-26, whereas only a few PKH-26–negative cells were observed in the center of the implanted collagen gels (Supplementary Fig. 3A). FACS analysis of enzymatically digested gels showed a decrease of PKH-26 fluorescence resulting from proliferation of cells, with cell divisions sequentially halving the initial fluorescence (Supplementary Fig. 3B). Notably, even after 8 weeks a minority of PKH-26–positive MSCs was proliferating (Ki67-positive) in the collagen gels (Supplementary Fig. 3C). We did not observe a significant foreign-body reaction of the host. Myeloperoxidase-positive granulocytes were only observed surrounding the suture used for fixation of the collagen gels, in terms of a small foreign body granulomas observed in every animal (Supplementary Fig. 4A). Staining for CD31 showed vascularization of implanted collagen gels. MSCs revealed a pericyte-like growth pattern surrounding CD31-positive endothelial cells (Supplementary Fig. 4B). Of note, MSCs were even able to secrete cytokines, as the stromal cell–derived factor 1 (SDF-1/CXCL12; Supplementary Fig. 4C), 8 weeks after transplantation.

Figure 3.

MSCs express Runx2 and sclerostin in both rat CKD models. (A) PKH-26–labeled MSCs (red) implanted in control rats did not express sclerostin or runx2, while MSC implanted in 5/6 nephrectomized rats (5/6Nx) and rats with adenine nephropathy (AdN) expressed sclerostin (B, arrows) and Runx2 (C, arrows). PKH-26–negative migrated cells did not express sclerostin or Runx2 (B, C, arrowheads). Scale bars = 20 µm, inserts 40 µm.

Upregulation of osteogenesis-related markers in the aortic wall and in the implanted MSC-containing collagen gels in CKD rats

Serum BMP-2 levels and parathyroid hormone (PTH) levels were, compared to the control group, significantly upregulated in both CKD rat models (Fig. 4A).

Figure 4.

Activation of osteogenic pathways in aortas and MSC-containing collagen gels of rat CKD models. (A) Serum levels of BMP-2 and PTH at 4 and 8 weeks. (B) Quantitative RT-PCR of BMP-2, Runx2, ALP, and SOST. Gene expression in the healthy control animals of every group (aorta, +MSC, –MSC) was set as 1. *p < 0.05, **p < 0.001.

Next we assessed osteogenesis marker in the aortic wall and in the implanted MSC-containing collagen gels. BMP-2 was nonsignificantly upregulated in the aortic wall and significantly in MSC-containing collagen gels of both CKD groups after 4 and 8 weeks in comparison to healthy controls (Fig. 4B). Runx2 was significantly upregulated in the aortic wall and +MSC gels in both CKD groups after 4 weeks. A significant upregulation of ALP was only detected in the MSC-containing collagen gels of the AdN group after 4 weeks, whereas after 8 weeks the ALP expression was significantly upregulated in both the aortic wall and the MSC-containing collagen gels of the AdN and the 5/6 Nx group. SOST expression was only significantly upregulated in the aortic wall of 5/6 Nx rats and in MSC-containing collagen gels in both CKD models after 8 weeks. There were no significant changes of the above mentioned markers in the non-MSC–containing collagen gels. Immunohistochemical analyses revealed expression of BMP-2 and sclerostin in the aortic wall and MSC-containing collagen gels of the CKD groups after 8 weeks, but not under control conditions (non-MSC collagen gels and healthy rats) (Supplementary Fig. 5).

Of note, vascular calcification is a process comparable to enchondral ossification of the bone. Thus, to analyze chondrogenic differentiation and a glycosaminoglycan-rich matrix, we performed Alcian blue staining and Sox9 quantitative realtime polymerase chain reaction (qt-RT-PCR) of the aortic wall as well as of collagen gels in control and CKD animals. As an indication of the enchondral ossification process, the matrix in both the aortic wall and the collagen gels with MSC in CKD animals was strongly Alcian blue–positive and Sox9 was significantly upregulated (Supplementary Fig. 6).

The pericyte marker 3G5 is expressed by primary MSC isolates and coexpressed with the osteocyte marker sclerostin during calcification

MSCs, in their role as vascular pericytes, express the pericytic marker 3G5 after isolation from the bone marrow (Fig. 5A). The expression of this pericytic marker is maintained in the implanted collagen gels. The implanted PKH-26–positive MSCs coexpress 3G5 and upon induction of CKD the osteocyte marker sclerostin (Fig. 5B). In the healthy, non-injured aortic wall of the control group, 3G5 is expressed in the adventitial layer. In contrast, during vascular calcification in both CKD models, cells adjacent to the calcified lesions in the media coexpress 3G5 and the osteocyte marker sclerostin (Fig. 5C).

Figure 5.

Expression of the pericyte marker 3G5 in implanted MSC and the aortic wall during calcification. (A) FACS analysis demonstrates expression of the pericyte marker 3G5 in isolated rat BM-MSCs. (B) Immunofluorescence staining of implanted PKH-26–labeled (red) MSCs for the pericyte marker 3G5 (green) and the osteocyte marker sclerostin (purple) indicating sclerostin expression in implanted 3G5-positive MSCs in rat models of vascular calcification. (C) In the control group the pericyte marker 3G5 is expressed only by cells in the adventitia of the aorta. (B, C) During vascular calcification of the aortic tunica media, 3G5-expressing cells are adjacent to the calcified area (asterisk) and coexpress the osteocyte marker sclerostin. A = adventitia; M = media; I = intima, * = calcified area. Scale bar: (A, B) 40 µm, (C insert) 20 µm.

CKD induces extensive ECM remodeling in the vascular wall and in MSC-containing collagen gels

The distribution pattern of ECM proteins was analyzed by immunohistochemistry and semiquantitatively analyzed in sections of all 8-week animals with a color-coded scoring system (Supplementary Fig. 7).

OPN is not expressed in the healthy aorta. The predominant collagen subtypes found in the healthy arterial wall are type I (70% to 75%) and type III (20% to 25%, mainly in the adventitia), with type IV comprising 1% to 2%.[22] Fibronectin is evenly distributed in the healthy vascular wall, predominantly in the media, whereas laminin expression is restricted to the adventitial tissue.[22]

Under control conditions, collagen gels with and without MSCs did not reveal OPN expression, were positive for collagen I (comprises 90% of the collagen matrix), showed slight collagen III/IV expression, and were negative for laminin (Supplementary Fig. 7).

In CKD, a pronounced change in the spatial distribution of ECM proteins occurred. In the arterial wall, the ECM of the media showed extensive remodeling processes and matrix deposits in close association to media calcification in both CKD models. OPN was significantly upregulated, aligning mainly media fibers and calcified areas. Collagen I and III are major components of the arterial wall; their expression was significantly upregulated in the media. Collagen IV and fibronectin were significantly upregulated in the CKD rat models. The strong fibronectin expression was not limited to the lamina media, but showed lamellar expression in the vascular wall, predominantly in association with vascular calcification. Laminin—mainly expressed in the adventitia under healthy conditions—also revealed a partly patchy, partly lamellar expression pattern in the arterial media of CKD rats (Supplementary Fig. 7).

Collagen gels containing MSCs revealed a similar pronounced upregulation of ECM proteins upon the induction of CKD (Supplementary Fig. 7). In contrast, collagen gels without MSCs showed no significant change in the ECM protein expression upon induction of CKD, suggesting that MSCs in the collagen gel acquired a pro-synthetic phenotype and were the driving force of the ECM remodeling and synthesis (Supplementary Fig. 7).

In line with the protein expression, qt-RT-PCR confirmed a significant upregulation of OPN, collagen I, collagen III, and fibronectin mRNA in the aortic wall and MSC-containing collagen gels in both CKD models. However, collagen IV was only significantly upregulated in the aortic wall of the AdN rats but in both CKD groups of the MSC-containing collagen gels. Laminin mRNA was only upregulated in the MSC-containing collagen gels of the 5/6 Nx group (Fig. 6).

Figure 6.

Gene expression data underlining similar matrix remodeling in the aortas and MSC-containing collagen gels of CKD rats after 4 and 8 weeks. Gene expression in the healthy control animals of every group (aorta, +MSC, –MSC) was set as 1. *p < 0.05, **p < 0.001.


Our current study analyses for the first time the effect of progressive CKD on heterotopic transplanted MSCs in vivo. We detected a number of interesting parallels between CKD-associated vascular calcification and differentiation of heterotopically implanted MSCs in CKD: (1) we describe osteogenic differentiation and calcification of MSCs; (2) we demonstrate an induction of similar osteogenic pathways in the vessel wall and in our heterotopic transplanted MSCs; and (3) we present data on matrix remodeling in implanted MSC-containing collagen gels comparable to aortic remodeling and sclerosis.

Our first major finding was the osteogenic differentiation and calcification of heterotopic transplanted MSCs occurring in parallel to vascular calcification in both CKD rat models. We demonstrate that MSCs not only calcified the surrounding collagenous matrix but upregulated reported markers of osteogenesis in parallel to mRNA expression in the aortic wall. Interestingly, the morphologist Alexander Friedenstein, in a historical publication—the first describing MSC as a fibroblast-like population in the BM—already observed osteogenesis in intraperitoneally transplanted diffusion chambers filled with BM cells, including bone fragments and small vessels, which prompted him to hypothesize an osteogenic stem cell population in the BM.[23]

BMP-2 is a candidate for inducing the osteogenic differentiation of mesenchymal cells and has been localized to atherosclerotic lesions.[24, 25] This potent embryonic differentiation factor of the transforming growth factor-β superfamily was recently discussed as an uremic toxin[26] and showed significantly upregulated serum levels in our CKD models, in line with previous data in human CKD patients by Dalfino and colleagues.[27] Interestingly, BMP-2 expression was significantly upregulated in the heterotopic transplanted collagen gels and showed a trend toward upregulation in the aortic wall in CKD.

The downstream targets of BMP-2, including Runx2, OPN, collagen I, and ALP, all previously described as playing a role in vascular calcification,[28-30] were also significantly upregulated in both the aortic wall and MSC-containing heterotopic collagen gels. Of note, both of our CKD rat models developed hyperphosphatemia due to CKD and a high-phosphorus diet. It is well known that hyperphosphatemia plays a major role in vascular calcification of CKD patients[31]; thus we have chosen rat models that reflect the hyperphosphatemic situation of CKD/ESRD patients. Our study cannot answer the question of whether uremia without hyperphosphatemia also has an impact on MSCs.

Sclerostin, a product of the SOST gene, is known as an osteocyte marker and indirect inhibitor of BMP signaling via inhibition of canonical Wnt signaling.[32, 33] Interestingly, BMP-2 is reported to increase sclerostin expression.[34] Recently, Zhu and colleagues[35] reported an upregulation of sclerostin in vascular smooth muscle cells during osteogenic differentiation in vitro and in the calcified media of Enpp–/– mice, indicating a role of sclerostin not only in bone formation but also in vascular calcification. We showed for the first time significantly upregulated aortic sclerostin expression in two rat models of CKD with vascular calcification and, importantly, also in MSC-containing heterotopic transplanted collagen gels. So far, the expression of SOST was detected in osteogenic differentiated MSCs and might reflect a negative feedback loop to control overwhelming calcification processes.[36]

In general, vascular structures are unique and provide a vascular matrix in close proximity to the endothelium, mainly consisting of collagen I. The vascular matrix is considered to be the “niche” or specialized microenvironment that determines the (stem) cell fate, cell polarity, and differentiation. We mimicked the vascular wall by the collagen I/III matrix of our collagen gels, providing both an important ECM protein of the vascular wall and regulator of cell differentiation.[12, 37] The ECM seems to be another important modulator of biomineralization because collagen type I in combination with ALP is necessary to induce osteogenic differentiation.[38] In both the aortic wall and MSC-containing collagen gels, collagen I is present under healthy conditions, but significantly upregulated under CKD conditions in combination with ALP, further highlighting the need for particular ECM proteins in calcification. It was recently described that these components of the calcified vascular wall are produced by vascular cells in vitro.[39, 40] This leads to our second major finding of a remodeling process with increased expression of ECM proteins in MSC-containing collagen gels of CKD rats, strongly resembling the vascular remodeling and sclerosis during vascular calcification of the arterial wall. In our opinion the process is most likely MSC-driven, because we observed (1) no significant changes in cell-free collagen gels, and (2) Runx2 expression in MSC, which is upstream of collagen I, OPN, ALP, and BMP-2. In line with prior studies showing upregulation of the ECM proteins OPN,[41, 42] collagen I/III/IV,[43, 44] fibronectin,[45] and laminin[46] in vascular calcification and sclerosis, we also observed an increased expression of these ECM proteins in the aortic wall of our CKD rats. To the best of our knowledge, we describe for the first time a very similar upregulation of these proteins in MSCs in an in vivo CKD environment. Remarkably, the upregulation of ECM proteins and osteogenic differentiation markers occurred in a comparable chronology in the aortic wall and in the heterotopic transplanted MSC-containing collagen gels, further highlighting the same differentiation pathways.

The aim of this study was to elucidate whether MSCs undergo osteogenic differentiation under a CKD vascular calcification environment in vivo. For definite proof of involvement of MSCs in the vascular calcification process, lineage-tracing experiments using a MSC-specific promoter are required. However, such a specific promoter has not yet been reported. Recent studies that used SM22-α Cre driver mouse lines for lineage tracing of VSMCs in LDLr–/– and ApoE–/– mice[47] and in MGP–/– mice[48] reported that the majority of calcifying cells in the vascular wall express SM22-α and concluded that smooth muscle cells are the major players in vascular calcification. However, it has also been reported that MSCs can gain SM22-α expression.[49, 50] Thus, the abovementioned lineage tracing cannot exclude the possibility of MSCs being involved in the calcification process of the arterial media.

In conclusion, we demonstrate an osteogenic differentiation and intensive ECM remodeling of heterotopic implanted MSCs in two rat models of CKD with vascular calcification. Our data indicate that MSCs undergo osteogenic differentiation in an in vivo setting of hyperphosphatemic CKD, supporting our hypothesis of MSCs as possible players in osteogenesis and ECM remodeling in soft-tissue calcification of CKD patients.


All authors state that they have no conflicts of interest.


This study was funded by intramural funding to RK, RKS, and UK (START grant) from RWTH Aachen University, by a grant from the German Society of Nephrology (DGFN) to RK, and grants from the European Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement HEALTH-F5-2008-223007 STAR-TREK and the German Research Foundation (Deutsche Forschungsgemeinschaft, SFB TRR57 P19) to JF and UK. We thank Esther Stuettgen (Department of Nephrology, RWTH Aachen University) and Norina Labude (Institute of Pathology, RWTH Aachen University) for excellent technical assistance in the laboratory.

Authors' roles: RK planned the project, executed the experiments, analyzed the data, wrote the manuscript, and put together the figures. UK, VMB, IL, JE, and BMK executed some experiments and contributed to data analysis and to writing of the manuscript. JF and RK contributed to writing of the manuscript. RKS planned the project, analyzed the data, wrote the manuscript, and put together the figures.